The present invention relates generally to pressure sensing transducers and pertains particularly to a package for transducers that is resistant to corrosive or conductive gasses and liquids.
Due to the hostile environment from highly corrosive fluids and the like, packages for electronic sensors measuring pressures in such environments are typically highly specialized, difficult to calibrate and expensive.
A pressure sensor (or pressure transducer) converts pressure to an electrical signal that can be easily measured. Sensors that incorporate micro-machining or MEMS (Micro-Electro-Mechanical System) technology are small and very accurate. Because they are fabricated similarly to the fabrication of commercial semiconductors they are also inexpensive to produce. FIG. 1 illustrates a MEMS pressure sensor 2 manufactured in accordance with the prior art. The topside 4 of the sensing element 6 (typically a silicon die) has defined resistors exhibiting a resistance that charges in magnitude in proportion to mechanical strain applied to die 6. Such resistors are called piezoresistive. The backside 8 of die 6 has a cavity 10 such that a thin diaphragm 12 of die material is formed. The alignment of the topside resistors and backside cavity 10 is such that the resistors are strategically placed in strain fields. When pressure is applied across diaphragm 12, diaphragm 12 flexes. The strain sensitive resistors and an associated circuit coupled thereto (not shown in FIG. 1) provide an electrical signal constituting a measure of this pressure.
Often, silicon die 6 is bonded to a support structure 14 with a bonding adhesive 15 or other method such as anodic bonding. Support structure 14, is bonded to a stainless steel plate 16 with a bonding adhesive 17. (Plate 16 is sometimes referred to as a header). Support structure 14 is made from a material such as glass or silicon, and helps isolate diaphragm 12 from sources of strain that are unrelated to pressure, e.g. thermal expansion or contraction of header 16. Support structure 14 includes a centrally defined opening 18 directly adjacent to and in fluid communication with cavity 10. Header 16 comprises a pressure port 19 in fluid communication with opening 18. This port 19 can be used to seal a vacuum in cavity 10. Alternatively, port 19 can be used to permit cavity 10 to be maintained at ambient pressure.
Header 16 is welded to a second port 20. Port 20 is connected to a body (e.g. a pipe, container or other chamber, not shown) containing fluid (e.g. a gas or a liquid) whose pressure is to be measured by sensor 2. Port 20 serves as a conduit for applying this fluid to sensor 2.
A drawback to MEMS sensors is that conductive and corrosive fluids (gases and liquids) can damage the sensor and the electronic structures (e.g. resistors) that are used to measure the pressure. Backside 8 of die 6 and adhesive bonds 15 and 17 are also susceptible to corrosion. To be used with corrosive or conductive fluids these sensors require some kind of isolation technique.
A popular isolation technique is to interpose a stainless steel diaphragm 22 between die 6 and port 20. Diaphragm 22 is welded to port 20 and header 16. A cavity 23 is thus formed between diaphragm 22 and header 16 and this cavity 23 is filled with a non-corrosive, non-conductive liquid such as silicone oil 24. Thus, diaphragm 22 and oil 24 isolate die 6 from any corrosive material in port 20.
When pressure is applied by the fluid in port 20 to diaphragm 22, diaphragm 22 deflects slightly, pressing on oil 24, which in turn presses on die 6. The pressure on die 6 is then detected by measuring the resistance of the piezoresistive resistors formed in diaphragm 12 of die 6. Corrosive media, the pressure of which is being measured, is kept away from the electronics by stainless steel diaphragm 22 and oil 24.
Header 16 often has at least one small hole 25 used to fill cavity 23 with oil 24. After cavity 23 is filled with oil 24, hole 25 is welded shut, e.g. with a welded ball 29. The design of FIG. 1 also includes metal pins 26 that are hermetically sealed to, but pass through, header 16. (Pins 26 are typically gold plated.) Gold or aluminum wires 28 are bonded to and electrically connect die 6 to metal pins 26. Pins 26 and wires 28 are used to connect die 6 to electronic circuitry (not shown in FIG. 1, but located below header 16) so that the resistance of resistors within die 6 can be measured.
A significant drawback the design of FIG. 1 is that when the temperature is increased, oil 24 expands and exerts pressure on stainless steel diaphragm 22 and sensor die 6. The resulting pressure change due to temperature causes the calibration of the sensor to change with temperature. The resulting errors introduced into the sensor measurements may contain linear and nonlinear components, and are hard to correct. The extent of this error is proportional to the amount of oil 24 contained in cavity 23. The more oil contained in cavity 23, the more oil there is to expand and thus more error over temperature. Currently existing designs require a substantial amount of oil for at least the following reasons: a) pressure sensing die 6 is enclosed inside oil filled cavity 23, and thus cavity 23 must be large enough to accommodate die 6; b) there are four hermetic pins 26 that must be wire bonded to die 6 (only two of which are shown in FIG. 1) so cavity 23 must also accommodate pins 26 and bonding wires 28; and c) cavity 23 must also accommodate manufacturing tolerances that are large enough to permit assembly of die 6, wiring 28 and the associated housing.
Another drawback to this design arises out of the fact that die 6 is made of silicon, which has a low coefficient of thermal expansion. Because die 6 must be mounted to stainless steel, and stainless steel has a relatively high coefficient of thermal expansion, a compliant die attach structure must be used. Typically this compliant die attach structure is a silicone elastomer. Because the silicone elastomers are not hermetic, when high vacuums are present, gas is in drawn through the silicone and into the oil. This causes large shifts in the offset calibration of the sensor due to the pressure of the gas drawn into cavity 23.
A third drawback to this design is the fact that hermetic feedthrough pins 26 are costly and problematic. In particular, this design requires metal pins 26 extending through glass regions 30 that serve as the hermetic seals. Glass 30 can crack. Also, pins 26 must be gold plated and flat on top to permit wire bonding. These designs are difficult to customize and the hermetic seals can be a leak point that must be checked before the sensor is assembled.
Attempts have been made to provide a corrosion resistant package using a non-fluid filled housing and polymeric or hermetic seals to seal the housing directly to the die. These methods allow corrosive material to travel inside and contact the die and sealing surfaces. Here, the amount of corrosion protection is limited because the sensor and associated seals are subject to damage by corrosive and possibly conductive materials. There have been some attempts to provide a polymeric barrier on the inside of the die and seal area. Conformal coatings such as Parylene or silicone materials only provide minimal corrosion improvement.
To maintain high quality and low cost it is desirable to construct an isolation technique that holds as little oil as possible, is readily assembled by automated processes, is easily modified for custom applications, and avoids unnecessary machining and assembly costs for hermetic feed through pins.